Midjourney V8 Launches with Dramatically Faster Image Generation Speeds, Accompanied by Steep Premium Pricing
Midjourney, the popular AI image generation platform, has unveiled version 8 of its core model, promising significant advancements in speed and quality. The rollout began quietly via the /settings command in the Midjourney Discord bot, catching many users by surprise. While the new model delivers up to five times faster generation compared to previous iterations, it introduces a controversial pricing structure that quadruples costs for the fastest modes, sparking debate among subscribers.
At the heart of V8 is a substantial performance boost. Standard generation times have been slashed, with images now producing in as little as 10 seconds under optimal conditions, a marked improvement over the roughly 50 seconds typical in V7. This acceleration stems from optimized inference processes and architectural refinements, enabling users to iterate designs more rapidly. Midjourney positions V8 as its most coherent and detailed model to date, excelling in rendering complex prompts with enhanced consistency across elements like anatomy, lighting, and textures.
Several new parameters enhance creative control. The image weight (–iw) slider now ranges from 0 to 5, allowing finer tuning of how much reference images influence outputs compared to text prompts. At 0, the model ignores images entirely, prioritizing textual descriptions; at maximum, it adheres closely to visual inputs. Character reference (–cref) has been upgraded for better multi-character consistency, supporting up to five references simultaneously. Style reference (–sref) and character reference now share the same random seed system, ensuring uniform aesthetics across generations. Additionally, --stylize now caps at 1000, with values above 100 yielding subtle enhancements rather than drastic shifts.
V8 also introduces experimental features like Draft Mode, accessible via the /draft command. This ultra-fast mode generates lo-res images in mere seconds, ideal for quick ideation before committing to full renders. However, it demands the Turbo generation mode, which unlocks only for higher-tier subscribers.
Accessing these capabilities hinges on Midjourney’s subscription tiers, unchanged in base structure but with inflated costs for speed. The Basic plan ($10/month) offers 3.3 GPU hours with standard Relax mode speeds. Standard ($30/month) provides 15 GPU hours, Pro ($60/month) 30 hours with Stealth Mode, and Mega ($120/month) 60 hours. Unlimited plans exist but cap Relax mode at 40 concurrent jobs.
The real contention lies in Fast and Turbo modes. Fast mode, already twice the Relax speed, costs $4 per GPU minute across all plans. Turbo, now five times faster than Relax (and thus 2.5 times faster than Fast), skyrockets to $16 per GPU minute, a fourfold increase over Fast pricing. Relax mode remains unmetered on paid plans, but Fast and Turbo incur per-minute fees that can accumulate quickly during intensive sessions. For context, generating 200 images in Turbo mode might exhaust a Basic plan’s allotment in under an hour, followed by $80+ in overage charges.
Midjourney justifies the premiums by noting Turbo’s resource intensity, reserving it for power users. Relax mode suffices for casual creators, while Fast suits moderate workflows. Yet, community feedback highlights frustrations: V8’s baseline speed already impresses, making Turbo feel indulgent amid rising costs. Free trials persist for newcomers, limited to 25 images via /imagine, but Discord’s evolving dynamics may limit accessibility.
To use V8, simply append --v 8 to prompts or select it in /settings. Early testers report superior prompt adherence, reduced artifacts, and vibrant outputs, particularly in photorealism and surrealism. Character consistency shines in multi-panel comics, with --cref producing reliable likenesses across poses and angles. However, minor quirks persist, like occasional over-saturation or style drifts at extreme --stylize values.
Midjourney’s iterative approach continues, with V8 still in alpha. Expect refinements via Discord announcements, as the company favors community input over traditional changelogs. For those wary of Discord reliance, web alpha access offers a cleaner interface, though bot commands remain primary.
This release underscores Midjourney’s balancing act: pushing AI frontiers while monetizing compute-heavy features. As competitors like Flux and Stable Diffusion advance open-source alternatives, Midjourney’s closed ecosystem and premium pricing may test user loyalty. For now, V8 elevates the platform’s technical prowess, delivering professional-grade results at unprecedented speeds, albeit with a premium attached.
Gnoppix is the leading open-source AI Linux distribution and service provider. Since implementing AI in 2022, it has offered a fast, powerful, secure, and privacy-respecting open-source OS with both local and remote AI capabilities. The local AI operates offline, ensuring no data ever leaves your computer. Based on Debian Linux, Gnoppix is available with numerous privacy- and anonymity-enabled services free of charge.
What are your thoughts on this? I’d love to hear about your own experiences in the comments below.